Microsoft seem to think that asking programmers to understand databases is too hard. So they have come up with yet another automatic code generator (following in the footsteps of Access, Data Controls, Typed Datasets, ... apologies to any silver bullets I may have dodged in the last 15 years).

Unfortunately the barrage of criticism from programmers has fastened on entirely the wrong point (as you might expect from a bunch of twenty-something blogshites).

Here's what's really wrong with it.

I believe that what the average object bigot likes to say is that:

private void SetLinesShippable(SalesOrder so)
{
	foreach (OrderLine oln in so.OrderLines)
	{
		oln.Shippable = oln.Product.IsAvailable;
	}
}

is more natural than

DECLARE @OrderLineId int
DECLARE @Available bit
	
DECLARE OrderLineCursor CURSOR FOR
	SELECT PR.ProductId
	FROM   ORDER_LINE
	WHERE  OrderId = @OrderId
	FOR UPDATE OF Shippable
		
OPEN OrderLineCursor
	
FETCH NEXT FROM OrderLineCursor INTO @ProductId
	
WHILE @@fetch_status = 0
BEGIN
	SELECT @Available = Available
	FROM   PRODUCT
	WHERE  ProductId = @ProductId
		
	UPDATE ORDER_LINE
	SET    Shippable = @Available
	WHERE  CURRENT OF OrderLineCursor
		
	FETCH NEXT FROM OrderLineCursor INTO @ProductId
END
	
CLOSE OrderLineCursor DEALLOCATE OrderLineCursor

So it is. But

UPDATE ORDER_LINE
SET    Shippable = PRODUCT.Available
FROM   PRODUCT
WHERE  ORDER_LINE.OrderId = @OrderId
AND    ORDER_LINE.ProductId = PRODUCT.ProductID

beats either of them. And won't choke performance on anything over 100 items.

That's what they don't get.